multiview radial basis function: a new approach on nonlinear forecasting of chaotic dynamic systems
Name: Maryam Masnadi-Shirazi
Grad Year: 2018
The curse of dimensionality has long been a hurdle in the analysis of complex data in areas such as computational biology, ecology, econometrics and etc. In this work, we present a forecasting algorithm that exploits the dimensionality of data in a nonlinear autoregressive framework. The main idea is that the dynamics of a chaotic dynamical system consisting of multiple time-series can be reconstructed using a combinations of multiple variables. This nonlinear autoregressive algorithm uses attractors reconstructed from a combination of variables as the inputs of a neural network to predict the future. We show that our approach, multiview radial basis function (MV-RBF) provides better forecast skill than that of a model-free approach, multiview embedding (MVE), for simulated ecosystems and a mesocosm experiment. By taking advantage of dimensionality, we show that MV-RBF overcomes the shortcomings of noisy and short time-series.
Industry Application Area(s)
Control Systems | Internet, Networking, Systems